Discrete utterance speech recognition without time alignment
John E. Shore, D. Burton · IEEE Transactions on Information Theory · 1983
The results of a new method are presented for discrete utterance speech recognition. The method is based on rate-distortion speech coding (speech coding by vector quantization), minimum cross-entropy pattern classification, and information-theoretic spectral distortion measures. Separate vector quantization code books are designed from training sequences for each word in the recognition vocabulary. Inputs from outside the training sequence are classified by performing vector quantization and finding the code book that achieves the lowest average distortion per speech frame. The new method obviates time alignment. It achieves99percent accuracy for speaker-dependent recognition of a20-word vocabulary that includes the ten digits, with higher accuracy for recognition of the digit subset. For speaker-independent recognition, the method achieves88percent accuracy for the20-word vocabulary and95percent for the digit subset. Background of the method, detailed empirical results, and an analysis of computational requirements are presented.
Cited by 85 works. doi:10.1109/tit.1983.1056716
Authors
References
- Some methods for classification and analysis of multivariate observations
- An Algorithm for Vector Quantizer Design
- Least squares quantization in PCM
- Minimum prediction residual principle applied to speech recognition
- Dynamic programming algorithm optimization for spoken word recognition
- Speech coding based upon vector quantization
- The condensed nearest neighbor rule (Corresp.)
- Distortion measures for speech processing
- Axiomatic derivation of the principle of maximum entropy and the principle of minimum cross-entropy
- Information Theory and Statistics